A Gaussian-Based Planning Approach for Robust Dose-Escalated Stereotactic Body Proton Therapy

Proton plans are often very homogeneous with minimal hotspots compared with the prescribed dose, which may not be ideal for ablative stereotactic body radiation therapy (SBRT). A treatment planning approach for stereotactic body proton therapy (SBPT) is proposed to match the characteristics of photon SBRT dose distributions, namely a Gaussian-shaped curve with high centralized doses in the gross tumor volume (GTV) and rapid dose falloff in healthy tissues.

The semiautomated technique is derived from a desired maximum dose value, or hotspot. The GTV is divided into concentric shells, and treatment plans are optimized to drive the hotspot to the innermost shell while using the remaining target shells to shape falloff within the GTV. The algorithm is tested in a cohort of 10 patients who previously received SBPT to the liver at our institution. The Gaussian approach is compared with standard homogeneous SBPT planning for both dosimetric and deliverability results.

Achieved hotspots on Gaussian plans matched the prespecified hotspot within 2.2% on average. Gaussian-based planning led to a statistically significant increase in biologically effective dose of 20 Gy (α/β = 10) compared with the homogeneous plans. On nominal plans, healthy liver and gastrointestinal organs-at-risk and planning-risk-volumes showed no statistically significant differences, though Liver D_mean trended lower with the Gaussian approach (10.8 Gy vs 11.9 Gy; P = .01). In evaluating plan robustness, Gaussian plans had slightly higher dose spill outside of the planning tumor volume, but lower dose to gastrointestinal organs-at-risk. In a subset of 3 plans evaluated for deliverability, Gaussian plans took 10 seconds longer to deliver per field, and patient-specific quality assurance results passed institutional criteria.

A mathematically-driven treatment planning technique was developed and tested for generating liver SBPT plans. The proposed technique was derived to mirror photon SBRT dose distributions, allowing for enhanced harmonization of SBPT with historical SBRT clinical data and potentially allowing safe radiation dose escalation during SBPT.